Autonomous, Mission-Based Traffic Engineering

For the first time in thirty years, we must contend with a near-peer adversary that is intent on making space a warfighting domain. To meet this challenge, the Navy is developing new beyond line-of-sight (BLOS) communications systems that are far more resilient to adversarial action. To fully realize the potential of these new systems, however, the Navy must also modernize the Automated Digital Network System (ADNS). ADNS manages IP network traffic between Navy afloat units and the shore. Several shortcomings of the Navy’s current approach to Quality of Service (QoS) management in ADNS can be identified. First, a common static policy is applied to all data traffic from each subnet. Second, strong bandwidth guarantees are not provided, only percentages of link rate. Finally, there is a single set of policies applied to every surface ship in the fleet, regardless of its role and regardless of its mission. In this SBIR, Caliola will develop a network orchestration tool that generates dynamic, mission tailored ADNS configurations which ensure traffic priorities are aligned with the Commander’s Intent while respecting BLOS bandwidth allocations. We call our tool AssuredConductor. Prior to a mission, AssuredConductor uses combinatorial optimization and network emulation to plan ADNS configurations. During a mission, reinforcement learning is used to change those configurations in response to an evolving threat environment, an evolving Commander’s Intent, or evolving traffic requirements. In Phase I, we will demonstrate key components of the AssuredConductor architecture. In Phase II, we will develop and demonstrate a Technology Readiness Level 6 prototype of AssuredConductor that builds on the software architecture and human machine interface constructs of WAMPS. WAMPS is the mission planner for the Wideband Anti-jam Modem System (WAMS) that we are building for PMW/A 170.